Quantitative differences in synthetic gut microbial inoculums do not affect the final stabilized in vitro community compositions

Abstract

In vitro studies of synthetic gut microbial communities (SGMCs) can provide valuable insights into the ecological structure and function of gut microbiota. However, the importance of the quantitative composition of an SGMC inoculum and its effect on the eventual stable in vitro microbial community has not been studied. To address this, we constructed two 114-member SGMCs differing only in their quantitative composition-one reflecting the average human fecal microbiome and another mixed in equal proportions based on cell counts. We inoculated each in an automated anaerobic multi-stage in vitro gut fermentor simulating two different colonic conditions, mimicking proximal and distal colons. We replicated this setup with two different nutrient media, periodically sampled the cultures for 27 days, and profiled their microbiome compositions using 16S rRNA gene amplicon sequencing. While nutrient medium explained 36% of the variance in microbiome composition, initial inoculum composition failed to show a statistically significant effect. Under all four conditions, paired fecal and equal SGMC inoculums converged to reach stable community compositions resembling each other. Our results have broad implications for simplifying in vitro SGMC investigations. IMPORTANCE In vitro cultivation of synthetic gut microbial communities (SGMCs) can provide valuable insights into the ecological structure and function of gut microbiota. However, it is currently not known whether the quantitative composition of the initial inoculum can influence the eventual stable in vitro community structure. Hence, using two SGMC inoculums consisting of 114 unique species mixed in either equal proportions (Eq inoculum) or resembling proportions in an average human fecal microbiome (Fec inoculum), we show that initial inoculum compositions did not influence the final stable community structure in a multi-stage in vitro gut fermentor. Under two different nutrient media and two different colon conditions (proximal and distal), both Fec and Eq communities converged to resemble each other's community structure. Our results suggest that the time-consuming preparation of SGMC inoculums may not be needed and has broad implications for in vitro SGMC studies.

Keywords: SHIME; in vitro gut models; synthetic gut microbial community.

Fig 1

Principal coordinate analysis (PCoA) using Bray-Curtis dissimilarity shows the composition of two theoretical SGMCs used in this study and 5,084 healthy fecal samples from 23 different countries. Abbreviated names of countries denote the mean country-specific microbiome compositions (Table S1).

Fig 2

Stabilization and convergence of SGMCs. (A) Microbiome composition differences between consecutive time points (2-day intervals) show stabilization by D11 in all eight compartments (PC/DC compartments with Fec/Eq inoculums fed with BSC/SF media). (B) Convergence of Fec and Eq community compositions under matched medium and colon compartment combinations. Theoretical composition difference between Fec and Eq inoculums is shown for reference. PC compartment is denoted by orange points and lines, and DC compartment is denoted by green points and lines. Bray-Curtis dissimilarity was used as beta diversity measure.

Fig 3

Comparing individual relative abundances of species in the fecal and equal SGMCs. Top row compares theoretical and observed inoculums. Second and third rows compare SGMCs grown in BSC and SF media, respectively. Sum of squared differences for each plot are also listed.

Fig 4

Detection patterns of the original inoculum taxa under different in vitro conditions. Phylogenetic tree of the 105 resolvable taxa that were part of the original inoculum is shown. Taxa are colored by their phylum affiliation; outer rings show the frequency of their detection in the inoculums, SF-fed and BSC-fed PC/DC compartments, across all time points. Black cells in the outer rings represent lack of detection. Species that are not distinguishable by their 16S rRNA V4 variable region sequence have been merged (see Table S4).